Research Methods & Statistics Lesson 2
Mysteries of Life We have questions Why do people behave that way? Is global warming occurring? Will my cancer come back? To get answers, we need… Information Data Explanation analysis & interpretation ~
Research Methods & Statistics Integral relationship Must consider both during planning Research Methods How data *are collected What kind of data Statistics Analysis depends on data & how it is collected ~
Scientific Inquiry Observing relationships between variables Variable: any measurable characteristic can take on many different values Science is a way thinking & observing 1. systematic empiricism 2. production of public knowledge 3. examination of solvable problems Guided by theory ~
Theories Descriptions Of observations Explanations Relationships between observations Generate hypotheses testable predictions about relationships ~
The Research Process Systematic Empiricism Observe Explain Predict Test Science is cumulative Theories tested Supported or modified No single experiment conclusive ~
Theories Scientific theories are falsifiable Certain outcomes would disprove them Non-scientific: Salsa dancing is the most exciting form of dance. Scientific: Salsa dancing increases physical fitness ~
Scientific Validity Scientific conclusions About relationships b/n variables Validity Soundness, legitimacy, truth Internal validity About cause & effect External (ecological) validity About broad applicability ~ Example: validity: don’t use thermometer to measure pulse Reliability: checking blood pressure, pulse
How are data collected? 2 scientific approaches Same or similar statistical analysis But not same conclusions Correlational methods Observe co-occurrence of variables Naturalistic observation, case studies, archival research, surveys, etc. Experimental method Manipulate a variable observe effect on another variable ~
Experimental Variables Independent (IV) Predictor (or cause) Manipulated Dependent (DV) Outcome (or effect) Measured Extraneous variables Or confounding Might also affect outcome (DV) ~
The Experimental Method At least 2 variables: Independent (IV) & Dependent (DV) At least 2 groups (levels of IV) control group - no treatment experimental - receives treatment random assignment to groups Control confounding variables Which might also affect DV Weakens internal validity ~
Variation within an Experiment Systematic Variation due to manipulation of IV Difference between groups Unsystematic Individual differences Variation due to random or uncontrolled variables Potentially confounding variables ~
Variation within an Experiment
Internal Validity Legitimacy of conclusions about cause & effect High internal validity Confident that only changes in IV cause change in DV Low internal validity Confounding variables influence outcome ~ Example: validity: don’t use thermometer to measure pulse Reliability: checking blood pressure, pulse
Two Experimental Approaches Treatment Level of IV or group Between-groups designs AKA Between-subjects, independent groups Each group gets different level of IV Within-subjects designs AKA repeated measures, dependent groups Each individual experiences each level of IV ~
Randomization Important for validity Helps avoid bias Random sampling (or selection) Selection of participants for study Representative sample from population external validity Random assignment to condition (groups) Minimize biasing of groups internal validity ~
Correlational vs. Experimental Internal vs External validity Inverse relationship based on control Correlational? internal vs external Cannot determine causality Experimental internal vs external Establishes cause & effect relationships For useful conclusions need both ~
Planning Research Correlational or experimental research Research design Between-groups or within-subjects Operational definition of variables Data categorical or quantitative Statistical analysis Depends on all of the above ~